Search for: All records

Female students are underrepresented in Computer Science degree programs at US colleges. This problem has resisted interventions for decades and average enrollment rates are stubbornly hovering around 18%, several success stories at selected institutions notwithstanding. Solutions to this problem require bringing in more female students and once enrolled keeping them in the program. To achieve the latter, improvements have been attempted at the curricular and pedagogical level on one hand, and on the social and community building level on the other. This paper describes a low cost approach to building a sense of community and belonging by making a Women in Computing club more "official" using a "token of belonging." Membership is made more rewarding by the promise of a conference trip 

Though hackathons are successful in attracting large crowds, they may not be sufficiently effective for broadening participation in computing, because they lack appeal for underrepresented groups in computing, and for people with family and job obligations. We propose a contrasting model for creating interest in computing, by making coding a spectator sport. We present an experience report on the design and implementation of a coding tournament, including survey results that informed the design of the system along with post-event questionnaire data from participants, exploring their attitudes towards different coding events. We find that coding tournaments can be an effective and engaging alternative to hackathons, and that they can motivate some audience members to pursue more coding activities, and possibly even participate as competitors in future tournaments. 

As the problem of drug abuse intensifies in the U.S., many studies that primarily utilize social media data, such as postings on Twitter, to study drug abuse-related activities use machine learning as a powerful tool for text classification and filtering. However, given the wide range of topics of Twitter users, tweets related to drug abuse are rare in most of the datasets. This imbalanced data remains a major issue in building effective tweet classifiers, and is especially obvious for studies that include abuse-related slang terms. In this study, we approach this problem by designing an ensemble deep learning model that leverages both word-level and character-level features to classify abuse-related tweets. Experiments are reported on a Twitter dataset, where we can configure the percentages of the two classes (abuse vs. non-abuse) to simulate the data imbalance with different amplitudes. Results show that our ensemble deep learning models exhibit better performance than ensembles of traditional machine learning models, especially on heavily imbalanced datasets.